End To End Dialogue Modelling On Multiwoz 2 0
评估指标
BLEU
MultiWOZ (Inform)
MultiWOZ (Success)
评测结果
各个模型在此基准测试上的表现结果
比较表格
| 模型名称 | BLEU | MultiWOZ (Inform) | MultiWOZ (Success) |
|---|---|---|---|
| galaxy-a-generative-pre-trained-model-for | 20.5 | 94.4 | 85.3 |
| pretraining-the-noisy-channel-model-for-task | 20.6 | 86.9 | 76.2 |
| task-oriented-dialog-systems-that-consider | 18.6 | 76.3 | 60.4 |
| soloist-few-shot-task-oriented-dialog-with-a | 16.5 | 85.5 | 72.9 |
| augpt-dialogue-with-pre-trained-language | 17.2 | 90.2 | 75.5 |
| a-simple-language-model-for-task-oriented | 15.0 | 84.4 | 70.1 |